END OF AN ERA, FRACTALFORUMS.COM IS CONTINUED ON FRACTALFORUMS.ORG

g'day mate i'm more or less in your neck of the woods too, new zealand.

that zoom animation you posted is pretty cool, the isometric height drawing makes it considerably more interesting than the usual zoom animations one finds, nice work!

always cool to see more coding experiments

incidentally, i think i know how you're choosing your constants over time for that animation! i rendered a similar looping animation, available here: http://www.fractographer.com/wip/julia2_x264.avi (needs the x264 codec, to play it use ffdshow available here: http://x264.nl)

looking forward to more of your works!

Hi LionHeart and welcome to the forums!

You are our 300th member!    Sorry there isn't a prize but welcome anyway.

that's how i made my animation too actually there are ways to follow the boundary of the m-set more closely, but they're a bit of a mission to implement.

btw, why do you save animations as small 256 colour gif images, when a more modern codec allows for so much better quality/filesize?

the reason i suggest not using gif images is exactly because of its (absolutely terrible) file size / quality characteristics! for example, some of your animated gifs are around 3mb in size, that buys a hell of a lot of mpeg4 encoded (truecolour, high resolution, 25fps) video.

to make animations i also save out images in sequence (e.g. "frameXYZW.bmp", i don't know if it supports png images), then use virtualdub to compile an avi from them. this also allows you to use its high quality resampling filters for antialiasing - just render at a higher resolution; moreover, it's a doddle to set the frame rate, mess around with the video compression settings, add an audio soundtrack, etc. then you just save the avi and that's that. i used to have my own avi writing code to avoid using all the hard disk space for the uncompressed frames, but virtualdub is really convenient.

virtualdub: http://www.virtualdub.org/
ffdshow package (with state of the art x264 codec): http://x264.nl/


about tracing the mandelbrot set boundary more accurately, that's really just an idea i've had for a while, haven't actually tried it: essentially you just numerically estimate the gradient of your potential function f : C -> R (where f is z^2 - bailout for the last iterated z value) and follow it. what troubles me is that, well, usually it's a good idea to use a higher order differential equation solvers like runge-kutta methods, however they depend on the existence of higher derivatives which fractal curves obviously don't have! so one obviously needs a bit of smoothing, which the euler method does provide by ignoring high order terms in the local taylor expansion; to compensate for this one would take lots of small steps, but if we follow it too accurately we'll probably get caught in a chaotic iteration following this crazy fractal curve... so there's a lot of ugly numerical tuning involved. which is why i didn't pursue it

for gif animations i use adobe imageready, try the trial maybe?

really flexible compression options, and it's just nice to work with.

i don't recommend gif at all though, especially since you compress the animation twice! have you tried that x264 codec? if you render a nicely anti-aliased and motion-blurred set of frames, let's say 512x512, pack it with x264 and with your gif pipeline to compare quality / filesize differences... you can expect the x264 video to be about 100x better.